Description Usage Arguments Details Value Author(s) References See Also Examples

It classifies and aligns the peaks
stored in the GRanges object. The method applies the k-mean alignment algorithm
with shift of the peaks and distance based on the convex combination of the
*L^p* distances between the spline-smoothed peaks and their derivatives. The order *p*
can be one of *1*, *2* and *∞*.

1 2 3 4 5 |

`object` |
GRanges object of length |

`parallel` |
logical. If |

`num.cores` |
integer. If |

`n.clust` |
integer vector (or scalar). Number of clusters in which the data set
is divided (possibly one, if |

`seeds` |
vector. Indices of the initial centers of the clusters, needed to initialize the k-mean procedure. The k-mean alignment, like all the k-mean-like algorithms,
is dependent on the choice of the initial centers
of the clusters, and each initialization
of the seeds can generate slightly different results. The
values must be included in |

`shift.peak` |
logical. It indicates whether the alignment via a translation of the abscissae
is performed ( |

`weight` |
real. Weight
with |

`subsample.weight` |
integer value. Number of data points used
to define the |

`alpha` |
real value between 0 and 1. Value of the convex weight |

`p` |
integer value in {0, 1 , 2}. Order of the |

`t.max` |
real value. It tunes the maximum shift allowed. In particular the maximum shift at each iteration is computed as
and the optimum registration coefficient will be identified between - max_shift and
+ max_shift. range( |

`plot.graph.k` |
logical. If |

`verbose` |
logical. If |

`rescale` |
logical. If |

See [Sangalli et al., 2010] and the package vignette for the complete description of the algorithm. The algorithm is completely defined once we fix the family of the warping function for the alignment and the distance function. In this function we focus only on the specific case of

warping functions: shifts with integer coefficients

*h(t) = t + c,*with

*c*an integer value;distance: convex combination of the

*L^p*distance between data and derivatives. The distance between*f*and*g*is*d(f, g) = (1 - α) || f - g ||_p + α w || f' - g' ||_p*The choice of

*|| . ||_p*corresponds to the value of`p`

in input. In particular`p = 0`

stands for*||.||_L^∞*,`p = 1`

for*|| . ||_L^1*and`p = 2`

for*|| . ||_L^2*

the GRanges `object`

with new metadata columns:

if

`align`

is`TRUE`

or`NULL`

, i.e. the clustering with alignment is performed the following metadata columns are added:`cluster_shift`

: for each peak, a vector of length equal to the maximum number of chosen clusters, containing at each position*k*the label of the cluster the peak is assigned to, when the total number of clusters is*k*and alignment is performed during the clustering. If*k*is not present in the`n.clust`

vector, the corresponding value is`NA`

.`coef_shift`

: for each peak, a vector of length equal to the maximum number of chosen clusters, containing at each position*k*the shift coefficient assigned to the peak, when the total number of clusters is*k*and alignment is performed during clustering. If*k*is not present in the vector`n.clust`

the corresponding value is`NA`

.`dist_shift`

: for each peak, a vector of length equal to the maximum number of chosen clusters, containing at each position*k*the distance of the specific peak from the corresponding center of the cluster, when the total number of clusters is*k*and alignment is performed during clustering. If*k*is not present in the vector`n.clust`

the corresponding value is`NA`

.

if

`shift.peak`

is`FALSE`

or`NULL`

, i.e. clustering is performed without alignment, the following metadata columns are added:`cluster_NOshift`

: for each peak, a vector of length equal to the maximum number of chosen clusters, containing at each position*k*the label of the cluster the peak is assigned to, when the total number of clusters is*k*and no alignment is performed during clustering. If*k*is not present in the vector`n.clust`

the corresponding value is`NA`

.`dist_NOshift`

: for each peak, vector of length equal to the maximum number of chosen cluster, containing at each position*k*the distance of the peak from the corresponding center of the cluster , when the total number of clusters is*k*and no alignment is performed during clustering. If*k*is not present in the vector`n.clust`

the corresponding value is`NA`

.

Alice Parodi, Marco J. Morelli, Laura M. Sangalli, Piercesare Secchi, Simone Vantini

Sangalli, L. M., Secchi, P., Vantini, S. and Vitelli, V., 2010. K-mean alignment for curve clustering. Computational Statistics and Data Analysis, 54 1219 - 1233.

choose_k

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 | ```
# load the data
data(peaks)
# cluster and align the data as a
# function of the
# number of cluster (from 1 to 5)
# with and without alignment.
# The automathically generated plot
# can be used to detect the
# optimal number of clusters and the
# classification method to be used
# (with or without alignment)
clustered_peaks <- cluster_peak ( peaks.data.summit, parallel = FALSE ,
n.clust = 1:5, shift.peak = NULL,
weight = 1, alpha = 1, p = 2,
plot.graph.k = TRUE, verbose = TRUE )
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.